Association Analysis of the Polymorphism of Human Leukocyte Antigen-A, -B and -E Gene with Behcet’s Disease in Japanese Cohort Using Sequencing-Based Typing Method

Behcet’s disease (BD) is a chronic inflammatory autoimmune disease and strongly associated with human leukocyte antigen (HLA) gene such as HLA-B and -A. To determine the association of the HLA-A, HLA-B and HLA-E alleles with BD in Japanese cohort as well as other cohorts, we performed sequencing-based typing method in 382 BD cases and healthy controls. Then, we analyzed the association of these alleles with BD. Our results indicated that HLA-B*51 is strongly associated with BD in Japanese cohort (Allelic model’s P value and OR are 4.59x10-26 and 4.59, respectively. Dominant model’s P value and OR are 2.1x10-24 and 5.32, respectively), while HLA-A*26 is also significantly associated with BD (Allelic model’s P value and OR are 5.3x10-5 and 1.80, respectively. Dominant model’s P value and OR are 8.6x10-6 and 2.06, respectively.). HLA-E*01:01 was not observed to have any significant association with BD. We also obtained the available data sets of Chinese and Japanese genome sequence from public data base and then, analyzed each Linkage disequilibrium (LD) structure using SNPs in HLA-A, -B and –E region. The pattern of LD including HLA-E was different between two cohorts. Collectively, we suggested that HLA-B*51 and HLA-A*26 are associated with BD in Japanese cohort as well as another cohort and that HLA-E*01:01 is not associated with BD. Our data suggest that LD structure of the region including HLA-E depends on a type of ethnic group even among Asian ethnic group.


Introduction
Behcet's disease (BD) is a chronic inflammatory autoimmune disease characterized by recurrent oral and genital ulcers, skin legions and uveitis [1]. T helper 1 (Th1) associated cytokines such as interferon-gamma and interleukin-12 were enhanced in BD patients [2]. Indeed, many studies have revealed the dominance of Th1 in the development of BD [3]. Recently, it's been thought that another T cell subsets, Th17 plays an important role in the pathogenesis and development of BD [4]. The incidence of BD is clearly high in the population who lives in from Middle East to Asia, along the "Old Silk Road", and rare in the USA especially in African Americans [5,6].
The patient's family members have a predisposition of BD including genetic and environmental factors. Indeed, the incidence of BD is clearly higher in those peoples than in general peoples [7]. Thus, epidemiological evidence indicates genetic factors contribute to the pathogenesis of BD. Several association studies have reported that human leukocyte antigen (HLA)-B*51

Patients and controls
We used a cohort of 384 Japanese unrelated BD patients and ethnically matched 384 unrelated healthy controls. All of the BD patients were diagnosed according to standard criteria [19] by the Japan Bechet's disease Research Committee at the Yokohama City University, Hokkaido University, Kurume University, Yuasa Eye Clinic and Fujioka Eye Hospital. These are included in our previous study [20] and [16]. The study methodology compiled with the guidelines of the Declaration of Helsinki. The study details were explained to all case and controls before obtaining their consent for genetic screening.

Quality control filtering
At first, quality control (QC) filtering of genotype data sets was performed to avoid false positive finding. We excluded two cases and two controls whose missing genotypes rates were more than 5%. After that, we could not exclude any SNPs showing departure from Hardy-Weinberg equilibrium the data sets consisted of 382 BD cases and 382 healthy controls were available. The OC filtering was performed by PLINK [21].

Genotype analysis of the HLA variants
Genomic DNA was extracted from peripheral blood lymphocytes using QIAamp DNA Maxi Kit (Qiagen, Hilden, Germany) according to the manufacture's instructions. To determine haplotypes of each HLA gene, we did sequencingbased typing method using primers previously reported in [22][23][24]. Determined alleles were confirmed to be consistent with WHO Nomenclature 2006 ImMunoGeneTics/HLA Database (URL: http://www.ebi.ac.uk/ipd/imgt/hla/).

Haplotype analysis
We obtained data sets of Han Chinese in Beijing (CHB) and Japanese of Tokyo (JPT) genome from International Hap Map Project (URL: http://hapmap.ncbi.nlm.nih.gov) and then, visualized LD structure between two ethnic groups showing association signals Haploview [25]. D' and R 2 were used for measurement of LD. The significant threshold was |D'| > 0.8 or R 2 >0.5.

Statistical analysis
The statistical significance of the difference in each HLA allelic frequencies between BD patients and healthy controls was determined by the chi-square test and Fisher's exact test using PLINK. The levels of significance was set to that P value is less than 0.05, and odds ratios (OR) with 95% confidence intervals (CI) were calculated. We fit the data to three genetic models such as allelic, dominant and recessive.

Results and Discussion
To examine whether or not HLA-A, -B or -E are associated with BD in Japanese cohort, we performed sequencing-based typing using large amount of BD patient and healthy control. In coincidence with previous reports [8][9][10][11], HLA-B*51 was strongly associated with BD in Japanese cohort when we calculated in allelic model and dominant model ( Table 1, P values are 4.59x10 -26 and 2.1x10 -24 , respectively. OR are 4.59 and 5.32, respectively.), although we did not in recessive model. HLA-A*26 is also significantly associated with BD in both allelic model and dominant model ( Table 1, P values are 5.3x10 -5 and 8.6x10 -6 , respectively. OR are 1.80 and 2.06, respectively.). However, HLA-E*01:01was not shown to have any significant association with BD even in three genetic models (Table 1). Park et al. [14] reported that HLA-E*01:01, which is one of the two major haplotypes, is significantly associated with BD in Korean cohort (P values and OR are 2.0x10 -4 and 0.7, respectively). They also reported that the polymorphism of CD94/NKG2A, which are an inhibitory receptor complex binding to HLA-E, are associated with BD in Korean cohort [26]. As such different susceptibility may be dependent on different genetic background of each cohort. To examine that, using available data sets from public database of human genome sequence, International HapMap Project, we also analyzed the difference of LD structure between JPT and CHB in the genomic region covering HLA-A, -B and -E (Figure 1). Our result indicates that LD pattern of the region containing HLA-E gene is different between JPT ( Figure 1A & 1B) and CHB ( Figure 1C &  1D). The pattern of LD in JPT was more associated with HLA-A region than in CHB (Figure 1 light green and yellowish line). We hypothesized that susceptibility of HLA-E in JPT may be much higher than in CHB because JPT has stronger LD with HLA-A ( Figure 1 light green and yellowish line). Indeed, any HLA-A allele is not significantly associated with BD in Han Chinese cohort [27], while it's associated in Japanese cohort [13]. In contrast to HLA-A region, those of HLA-B were quite similar LD pattern between JPT and Chinese (Figure 1 Light blue line). These regions may be quite conserved among ethic groups. Taken together, we concluded that HLA-B*51 and -A*26 are associated with BD in Japanese cohort as well as another cohorts and that HLA-E*01:01 is not associated with BD in Japanese cohort although previous